The Assurance Desk
Collect only the personal data you need, and keep safety in view when using AI. Those are the themes in selected posts from Daniel Solove and Lisa Forte.
Daniel Solove: is collecting less data working?
Solove asks why data minimisation is falling short. That is his view, rather than a finding measured by this round-up.
Data minimisation means keeping enough personal data for your purpose, but no more than you need. The ICO says to start by being clear about that purpose.
Data minimization is the key to the success of privacy laws. Why is it working so badly?
Lisa Forte: keep safety alongside innovation
Forte argues that innovation should not come at the cost of safety or security. Her LinkedIn post makes a case for how AI should be used. It does not measure how safe a particular product is.
This is a useful question for teams building AI and teams using it. What could go wrong, and what would you change before going ahead?
Innovation can’t be accepted at the cost of safety or security
Turn the discussion into useful questions
The ICO says to review the personal data you hold and delete what you no longer need. Its AI guidance also covers accountability, impact assessments and transparency, meaning how people understand what is happening to their data.
| Theme | A question to ask at work | The guidance behind it |
|---|---|---|
| Purpose | Why do we need each piece of personal data? | Be clear about the purpose before deciding how much data to hold. |
| Review | What do we hold that we no longer need? | Review the data and delete what is no longer needed. |
| AI oversight | Who will assess the risks and explain how the tool uses data? | The AI guidance covers accountability, impact assessments and transparency. |
Read the guidance with its current status in mind
The ICO marks its AI and data protection guidance as under review following changes made by the UK Data (Use and Access) Act. Treat it as background, and check the current rules before relying on a particular passage.
The guides below explain the data protection principles, AI impact assessments and approvals for AI agents. Use the one that matches the work you are doing.
Read next
Sources
- Daniel Solove on Bluesky, @daniel-solove.bsky.social on Bluesky, checked 13 October 2026
- Lisa Forte on LinkedIn, Opening twitter this morning and seeing the next dose of AI nonsense being shared everywhere. Even I find myself becoming desensitised to reading how all these bad things are going to happen. Now…, Lisa Forte, 25 comments, checked 13 October 2026
- ICO, Principle (c): Data minimisation, checked 13 October 2026
- ICO, Guidance on AI and data protection, checked 13 October 2026
This is an independent editorial guide, not legal advice.